CN112435333A - Road scene generation method and related device - Google Patents
Road scene generation method and related device Download PDFInfo
- Publication number
- CN112435333A CN112435333A CN202011095861.4A CN202011095861A CN112435333A CN 112435333 A CN112435333 A CN 112435333A CN 202011095861 A CN202011095861 A CN 202011095861A CN 112435333 A CN112435333 A CN 112435333A
- Authority
- CN
- China
- Prior art keywords
- target
- road
- information
- instance
- point cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Geometry (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Remote Sensing (AREA)
- Computer Graphics (AREA)
- Traffic Control Systems (AREA)
Abstract
The application discloses a road scene generation method and a related device, and relates to an artificial intelligent automatic driving technology. Acquiring target point cloud data corresponding to a target road; then extracting key features in the target point cloud data; clustering the key features according to the feature identification to obtain a target instance and at least one boundary instance; establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and a related apparatus for generating a road scene.
Background
The automatic driving technology comprises the technologies of high-precision maps, environment perception, behavior decision, path planning, motion control and the like, and has wide application prospects. In the process of testing the automatic driving, the credibility of the automatic driving simulation result is closely related to whether the simulation environment is close to the real environment or not.
The process of automatic driving simulation relates to the simulation of a road scene, for example, for highway pavement reconstruction, rough reconstruction of a highway on a macroscopic level is carried out mainly through sparse data such as field measurement data of a total station, road design data or characteristic data extracted by images so as to obtain the approximate position and the trend of the highway, and then artificial modeling is carried out based on the position and the trend so as to obtain the road scene.
However, the sparse data used by the highway reconstruction technology cannot provide information consistent with the real environment, and the time and labor are consumed in the process of manual modeling, so that errors are easily introduced, the road scene of the automatic driving simulation is inconsistent with the road scene of the real environment, and the accuracy of the road scene is affected.
Disclosure of Invention
In view of this, the present application provides a method for generating a road scene, which can generate the road scene with high accuracy.
A first aspect of the present application provides a method for generating a road scene, which may be applied to a system or a program that includes a function of generating a road scene in a terminal device, and specifically includes:
collecting target point cloud data corresponding to a target road;
extracting key features in the target point cloud data, wherein the key features correspond to feature identifications;
clustering the key features according to the feature identification to obtain a target instance and at least one boundary instance, wherein the target instance is used for indicating the backbone information corresponding to the target road, and the boundary instance is used for indicating the range information corresponding to the target road;
establishing a target space based on the backbone information and the range information;
and inserting detail features into the target space to generate a target road scene.
Optionally, in some possible implementation manners of the present application, the acquiring point cloud data corresponding to a target road includes:
collecting segmented point cloud data corresponding to each moment of the target road in a preset time period;
converting the segmented point cloud data into a coordinate system of a target moment according to a transformation matrix to obtain converted point cloud data;
and superposing the converted point cloud data to obtain the target point cloud data.
Optionally, in some possible implementations of the present application, the extracting key features in the target point cloud data includes:
calling a target neural network based on the target point cloud data, wherein the target neural network is obtained by training based on training data containing training characteristics, the training characteristics comprise the characteristic identification, and the training data is related to the target point cloud data;
inputting the target point cloud data into a target neural network to extract the key features.
Optionally, in some possible implementation manners of the present application, the clustering the key features according to the feature identifier to obtain a target instance and at least one boundary instance includes:
segmenting the key features based on the feature identification to obtain a target feature set and at least one boundary feature set;
connecting the features contained in the target feature set to obtain the target instance;
and connecting the features contained in the boundary feature set to obtain at least one boundary instance.
Optionally, in some possible implementations of the present application, the connecting features included in the target feature set to obtain the target instance includes:
connecting the features contained in the target feature set to obtain a route example;
and performing curve fitting on the route example to obtain the target example.
Optionally, in some possible implementations of the present application, the performing curve fitting on the route instance to obtain the target instance includes:
extracting discrete point groups from the route instance;
pairing discrete points in the discrete point group to determine corresponding direction vectors;
determining control points based on the direction vectors to determine fitting parameters;
and performing curve fitting according to the fitting parameters to obtain the target example.
Optionally, in some possible implementations of the present application, the method further includes:
determining fluctuation information of adjacent points in the route example;
screening the fluctuation information based on a simplified algorithm to update the adjacent point locations;
adjusting the route instance based on the updated neighboring point locations.
Optionally, in some possible implementations of the present application, the method further includes:
acquiring start and stop point information corresponding to the route example;
determining a target connecting line based on the start and stop point information;
and updating the route example according to the intersection condition of the target connecting line and the route tangent, wherein the route tangent is determined based on the route example.
Optionally, in some possible implementation manners of the present application, the establishing a target space based on the backbone information and the range information includes:
determining a candidate range based on the backbone information;
and updating the candidate range according to the range information to establish the target space.
Optionally, in some possible implementation manners of the present application, the updating the candidate range according to the range information to establish the target space includes:
determining a local variation region indicated in the range information;
determining distance information of the local variation region and the candidate range based on ray casting;
and updating the candidate range according to the distance information to establish the target space.
Optionally, in some possible implementations of the present application, the inserting the detail feature into the target space to generate the target road scene includes:
determining width information corresponding to the target road according to the backbone information and the range information;
establishing a target grid in the target space according to the width information and the backbone information;
invoking a detailed feature mapping into the target mesh to generate the target road scene.
Optionally, in some possible implementations of the present application, the method further includes:
acquiring the running record information of the target road;
generating a driving texture feature based on the driving record information;
and updating the target road scene according to the driving texture features.
A second aspect of the present application provides a device for generating a road scene, including: the acquisition unit is used for acquiring target point cloud data corresponding to a target road;
the extraction unit is used for extracting key features in the target point cloud data, and the key features correspond to feature identifications;
the clustering unit is used for clustering the key features according to the feature identifiers to obtain a target instance and at least one boundary instance, wherein the target instance is used for indicating the backbone information corresponding to the target road, and the boundary instance is used for indicating the range information corresponding to the target road;
a generating unit configured to establish a target space based on the backbone information and the range information;
the generating unit is further configured to insert a detail feature into the target space to generate a target road scene.
Optionally, in some possible implementation manners of the present application, the collecting unit is specifically configured to collect segmented point cloud data corresponding to each time of the target road within a preset time period;
the acquisition unit is specifically used for converting the segmented point cloud data into a coordinate system of a target moment according to a transformation matrix to obtain converted point cloud data;
the acquisition unit is specifically configured to perform superposition based on the converted point cloud data to obtain the target point cloud data.
Optionally, in some possible implementations of the present application, the extracting unit is specifically configured to invoke a target neural network based on the target point cloud data, where the target neural network is trained based on training data including training features, the training features include the feature identifiers, and the training data is related to the target point cloud data;
the extraction unit is specifically configured to input the target point cloud data into a target neural network to extract the key features.
Optionally, in some possible implementation manners of the present application, the clustering unit is specifically configured to segment the key feature based on the feature identifier to obtain a target feature set and at least one boundary feature set;
the clustering unit is specifically configured to connect features included in the target feature set to obtain the target instance;
the clustering unit is specifically configured to connect features included in the boundary feature set to obtain at least one of the boundary instances.
Optionally, in some possible implementation manners of the present application, the clustering unit is specifically configured to connect features included in the target feature set to obtain a route instance;
the clustering unit is specifically configured to perform curve fitting on the route instance to obtain the target instance.
Optionally, in some possible implementations of the present application, the clustering unit is specifically configured to extract discrete point groups from the route instances;
the clustering unit is specifically configured to pair the discrete points in the discrete point group to determine corresponding direction vectors;
the clustering unit is specifically configured to determine control points based on the direction vectors to determine fitting parameters;
and the clustering unit is specifically used for performing curve fitting according to the fitting parameters to obtain the target example.
Optionally, in some possible implementation manners of the present application, the clustering unit is specifically configured to determine fluctuation information of adjacent point locations in the route instance;
the clustering unit is specifically configured to screen the fluctuation information based on a simplified algorithm to update the adjacent point locations;
the clustering unit is specifically configured to adjust the route instance based on the updated adjacent point location.
Optionally, in some possible implementation manners of the present application, the clustering unit is specifically configured to obtain start point and stop point information corresponding to the route instance;
the clustering unit is specifically used for determining a target connecting line based on the start and stop point information;
the clustering unit is specifically configured to update the route instance according to an intersection condition of the target connecting line and a route tangent, where the route tangent is determined based on the route instance.
Optionally, in some possible implementations of the present application, the generating unit is specifically configured to determine a candidate range based on the backbone information;
the generating unit is specifically configured to update the candidate range according to the range information to establish the target space.
Optionally, in some possible implementations of the present application, the generating unit is specifically configured to determine a local variation region indicated in the range information;
the generating unit is specifically configured to determine distance information between the local variation region and the candidate range based on ray casting;
the generating unit is specifically configured to update the candidate range according to the distance information to establish the target space.
Optionally, in some possible implementation manners of the present application, the generating unit is specifically configured to determine width information corresponding to a target road according to the backbone information and the range information;
the generating unit is specifically configured to establish a target grid in the target space according to the width information and the backbone information;
the generating unit is specifically configured to invoke a detailed feature to be mapped into the target grid to generate the target road scene.
Optionally, in some possible implementation manners of the present application, the generating unit is specifically configured to obtain driving record information of the target road;
the generating unit is specifically used for generating driving texture features based on the driving record information;
the generating unit is specifically configured to update the target road scene according to the driving texture feature.
A third aspect of the present application provides a computer device comprising: a memory, a processor, and a bus system; the memory is used for storing program codes; the processor is configured to execute the method for generating a road scene according to any one of the above first aspect or the first aspect according to instructions in the program code.
A fourth aspect of the present application provides a computer-readable storage medium, having stored therein instructions, which, when run on a computer, cause the computer to execute the method for generating a road scene according to the first aspect or any one of the first aspects.
According to an aspect of the application, a computer program product or computer program is provided, comprising computer instructions, the computer instructions being stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method for generating a road scene provided in the first aspect or the various optional implementations of the first aspect.
According to the technical scheme, the embodiment of the application has the following advantages:
acquiring target point cloud data corresponding to a target road; then extracting key features in the target point cloud data, wherein the key features correspond to the feature identifiers; clustering the key features according to the feature identification to obtain a target example and at least one boundary example, wherein the target example is used for indicating the backbone information corresponding to the target road, and the boundary example is used for indicating the range information corresponding to the target road; further establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a network architecture diagram of a road scene generation system operation;
fig. 2 is a flowchart of generating a road scene according to an embodiment of the present disclosure;
fig. 3 is a flowchart of a method for generating a road scene according to an embodiment of the present disclosure;
fig. 4 is a scene schematic diagram of a method for generating a road scene according to an embodiment of the present application;
fig. 5 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 6 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 7 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 8 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 9 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 10 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 11 is a flowchart of another road scene generation method according to an embodiment of the present application;
fig. 12 is a scene schematic diagram of another road scene generation method according to an embodiment of the present application;
fig. 13 is a schematic structural diagram of a device for generating a road scene according to an embodiment of the present disclosure;
fig. 14 is a schematic structural diagram of a terminal device according to an embodiment of the present application;
fig. 15 is a schematic structural diagram of a server according to an embodiment of the present application.
Detailed Description
The embodiment of the application provides a road scene generation method and a related device, which can be applied to a system or a program containing a road scene generation function in terminal equipment, and target point cloud data corresponding to a target road are collected; then extracting key features in the target point cloud data, wherein the key features correspond to the feature identifiers; clustering the key features according to the feature identification to obtain a target example and at least one boundary example, wherein the target example is used for indicating the backbone information corresponding to the target road, and the boundary example is used for indicating the range information corresponding to the target road; further establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims of the present application and in the drawings described above, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are, for example, capable of operation in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "corresponding" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
First, some nouns that may appear in the embodiments of the present application are explained.
Point cloud: and obtaining a characteristic set after acquiring the space coordinate of each sampling point on the surface of the object.
Road scene: three-dimensional scenes for autopilot simulation.
A semantic segmentation model: a neural network model combining image classification, target detection and image segmentation is characterized in that an image is segmented into region blocks with certain semantic meanings, the semantic category of each region block is identified, the semantic reasoning process from a bottom layer to a high layer is achieved, and finally a segmented image with pixel-by-pixel semantic labels is obtained.
Douglas-pock algorithm (Douglas-Peucker algorithm): an algorithm that approximates a curve as a series of points and reduces the number of points.
It should be understood that the method for generating a road scene provided by the present application may be applied to a system or a program including a function of generating a road scene in a terminal device, for example, an interactive drama, specifically, the system for generating a road scene may operate in a network architecture as shown in fig. 1, which is a network architecture diagram in which the system for generating a road scene operates as shown in fig. 1, and as can be seen from the figure, the system for generating a road scene may provide a process of generating a road scene with multiple information sources, that is, point cloud data is collected by different information sources, and is sent to a server to extract features and generate a feature instance, so that a corresponding road scene may be generated and sent to the corresponding information source, so as to facilitate the information source to perform automatic driving simulation based on the road scene; it can be understood that, fig. 1 shows various terminal devices, in an actual scene, there may be more or fewer types of terminal devices participating in the process of generating a road scene, and the specific number and types depend on the actual scene, which is not limited herein, and in addition, fig. 1 shows one server, but in an actual scene, there may also be participation of multiple servers, especially in a scene with multiple capturing devices, the specific number of servers depends on the actual scene.
In this embodiment, the server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a network service, cloud communication, a middleware service, a domain name service, a security service, a CDN, a big data and artificial intelligence platform, and the like. The terminal may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and the like. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein.
It is understood that the generation system of the road scene may be operated in a personal mobile terminal, for example: the application, such as an interactive play, can also be operated on a server, and can also be operated on a third-party device to provide generation of a road scene so as to obtain a generation processing result of the road scene of the information source; the specific road scene generation system may be operated in the above-mentioned device in the form of a program, may also be operated as a system component in the above-mentioned device, and may also be used as one of cloud service programs, and the specific operation mode is determined by an actual scene, and is not limited herein.
The automatic driving technology generally comprises technologies such as high-precision maps, environment perception, behavior decision, path planning, motion control and the like, and has wide application prospects. In the process of testing the automatic driving, the credibility of the automatic driving simulation result is closely related to whether the simulation environment is close to the real environment or not.
The road model generation process in the automatic driving simulation process relates to Computer Vision technology (CV), and Computer Vision calculation is a science for researching how to make a machine see, and further means that a camera and a Computer are used for replacing human eyes to perform machine Vision such as identification, tracking and measurement on a target, and further performing image processing, so that the Computer processing becomes an image more suitable for human eyes to observe or is transmitted to an instrument to detect. As a scientific discipline, computer vision research-related theories and techniques attempt to build artificial intelligence systems that can capture information from images or multidimensional data. Computer vision technologies generally include image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content/behavior recognition, three-dimensional object reconstruction, 3D technologies, virtual reality, augmented reality, synchronous positioning, map construction, and other technologies, and also include common biometric technologies such as face recognition and fingerprint recognition.
The automated driving simulation process may also involve machine learning. Machine Learning (ML) is a multi-domain cross discipline, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning.
The process of automatic driving simulation relates to the simulation of a road scene, for example, for highway pavement reconstruction, rough reconstruction of a highway on a macroscopic level is carried out mainly through sparse data such as field measurement data of a total station, road design data or characteristic data extracted by images so as to obtain the approximate position and the trend of the highway, and then artificial modeling is carried out based on the position and the trend so as to obtain the road scene.
However, the sparse data used by the highway reconstruction technology cannot provide information consistent with the real environment, and the time and labor are consumed in the process of manual modeling, so that errors are easily introduced, the road scene of the automatic driving simulation is inconsistent with the road scene of the real environment, and the accuracy of the road scene is affected.
In order to solve the above problem, the present application proposes a method for generating a road scene, which is applied to the flow framework of the generation of the road scene shown in fig. 2, as shown in fig. 2, for the flow architecture diagram generated in the road scene provided in the embodiment of the present application, a control instruction is first sent at the interface layer, and the collection vehicle carrying the lidar collects road surface environment data in the target highway, then the road surface is sent to an interface layer to establish the corresponding relation between the real road surface environment and the digital road surface environment, key elements for reconstructing the highway road surface are extracted, then a trunk road, a fork road and a locally widened road surface of the highway are reconstructed from the extracted key elements of the road surface by a digital geometric calculation method, and automatically generating a three-dimensional grid model and texture coordinates of the road surface, thereby generating a road scene for automatic driving simulation.
It can be understood that the method provided by the present application may be a program written as a processing logic in a hardware system, or may be a road scene generation device, and the processing logic is implemented in an integrated or external manner. As an implementation manner, the generation device of the road scene acquires target point cloud data corresponding to a target road; then extracting key features in the target point cloud data, wherein the key features correspond to the feature identifiers; clustering the key features according to the feature identification to obtain a target example and at least one boundary example, wherein the target example is used for indicating the backbone information corresponding to the target road, and the boundary example is used for indicating the range information corresponding to the target road; further establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
The scheme provided by the embodiment of the application relates to an artificial intelligence automatic driving technology and a computer vision technology, and is specifically explained by the following embodiments:
with reference to the above flow architecture, a method for generating a road scene in the present application will be described below, please refer to fig. 3, where fig. 3 is a flow chart of a method for generating a road scene provided in an embodiment of the present application, where the management method may be executed by a terminal device, or by a server, or by both the terminal device and the server, and the following description will be given by taking an example of execution of the terminal device. The embodiment of the application at least comprises the following steps:
301. and collecting target point cloud data corresponding to the target road.
In this embodiment, the target road may be an expressway or a road with specific identification features, and the specific acquisition process may be directed to a part of a section of the specified target road or the entire road; the target point cloud data may be obtained by cooperation of a plurality of collection devices, for example, a data collection vehicle equipped with a laser radar (Lidar), an Inertial Measurement Unit (IMU), and a Global Navigation Satellite System (GNSS) to collect data on a target road.
Specifically, the generation process of the target point cloud data may be a data superposition process performed at multiple acquisition time points, that is, segmented point cloud data corresponding to a target road at each moment in a preset time period is acquired at first; then, converting the segmented point cloud data into a coordinate system of a target moment according to the transformation matrix to obtain converted point cloud data; and then, the target point cloud data is obtained by superposing the converted point cloud data, so that the accuracy of the target point cloud data is ensured.
In one possible scenario, the determination of the transformation matrix may be calculated by the acquisition device, for example assuming that the vehicle is located at time t in a position ot(x, y, z), the point cloud obtained by the laser radar is Pt. The data acquisition vehicle can obtain the time t of the vehicle from any data acquisition through the IMU and the GNSSaTo tbIs transformed bySatisfy the requirement ofWith the aid of a transformation matrix from a pre-calibrated lidar sensor center to a collection vehicle center, the collection vehicle is assisted by a series of collection times T ═ T (T)1,t2,…,tN) Acquired transformed point cloud data (P)1,P2,…,PN) Conversion to a starting instant t by means of the above-mentioned transformation matrix1Is (P) 'in the coordinate system of (A)'1,P′2,…,P′N) Obtaining target point cloud information P ═ P 'of the collected target highway environment'1+P′2+…+P′NThereby ensuring the integrity of the target point cloud information。
302. And extracting key features in the target point cloud data.
In this embodiment, the key feature corresponds to the feature identifier, where the key feature is a feature used to indicate a road direction and a boundary, for example, in an expressway, a light barrier is generally disposed on a road centerline, so that the light barrier can indicate the direction of the expressway, that is, the light barrier is a key feature in the target point cloud data; in addition, for key features that indicate boundaries, include curbs, median strips, and the like.
In a specific scenario, the determination of the key features may be determined based on a design standard of a highway road surface, for example, according to "highway engineering technology standard [ enclosed article description ] JTGB 01-2014", the traffic volume of general highway project design and service is bidirectional and equivalent, and except for special areas such as locally arranged auxiliary lanes, acceleration and deceleration lanes, emergency stop belts, danger avoiding lanes, climbing lanes and the like, other areas of the general highway are symmetrically designed along a roadbed cross section. Specifically, as shown in fig. 4, fig. 4 is a scene schematic diagram of a method for generating a road scene provided in the embodiment of the present application. The figures show the design criteria of various lanes on a highway, wherein the middle zone is located in the center of the road, i.e. can be used as a key feature for indicating the road direction; the hard road shoulder and the soil road shoulder can indicate the road boundary of the expressway, namely can be used as key features for indicating the road boundary; in addition, the final road may be obtained by using the boundary lines in the lanes to recognize the respective lanes.
Optionally, the extraction process of the key features may be performed based on a deep neural network, that is, a target neural network is called based on target point cloud data, where the target neural network is trained based on training data including training features, the training features include the feature identifiers, and the training data is related to the target point cloud data, for example, the training data may be an image set having similar key features to the target point cloud data; thus, the target point cloud data can be input into the target neural network to extract key features. The accuracy of extracting the key features is guaranteed by carrying out targeted training on the target neural network.
In a possible scene, setting key features as a road edge stone (boundary), an isolation zone (boundary) and a light barrier (central line), labeling training data based on image information of the key features, and then training a semantic segmentation model in a target neural network by using the labeled training data, so that the target neural network has the capability of extracting key road surface feature information. Therefore, the accuracy of extracting the key features from the target point cloud data is ensured.
It is understood that, in the above embodiments, three object-assisted reconstructions, namely, a road edge stone, an isolation zone, and a light barrier, are used as an example, elements having similar functions to these road elements may also be used as key features, such as an isolation net (central line), a road rod (boundary), and a road sign (boundary), and in addition, for the number of key feature types, it is necessary to include key features indicating a road direction dimension and key features indicating a road boundary, that is, at least two or more types, and the specific feature form and the number of types are determined by an actual scene, and are defined herein.
303. And clustering the key features according to the feature identification to obtain a target instance and at least one boundary instance.
In this embodiment, the target instance is used to indicate the backbone information corresponding to the target road, and the boundary instance is used to indicate the range information corresponding to the target road; the backbone information, i.e., the trend context of the target road, may be the center line of the corresponding target instance, the range information, i.e., the boundary range of the target road, may be the road edge stone of the corresponding boundary instance, and the specific example type is determined by the actual scene.
It can be understood that, because similar feature identifiers have similar features, road elements of the same category can be obtained after clustering key features according to the feature identifiers, for example, all the road edge stones are labeled as one category. Specifically, in the clustering process, the road surface feature information obtained by semantic segmentation is clustered in an Euclidean space, and the example segmentation result of the key features is further obtained, so that a target example and at least one boundary example are obtained. When the boundary example is one, namely the road is a single lane, the target example is a boundary line of the road, and the boundary example is also the boundary line of the road; and when the number of the boundary instances is two, as shown in fig. 5, fig. 5 is a scene schematic diagram of another road scene generation method provided in the embodiment of the present application. The figure shows a target example A1, namely a central line, and a boundary example A2 and a boundary example A3, namely a road edge stone, so that the trend and the range of a target road can be described; further, when there are more boundary instances, it may be a recognition process for further lanes of the road, such as recognizing the edge of a tidal vehicle as a boundary instance, etc.
Specifically, the generation process of the instance may be a process of connecting a plurality of features, that is, the key features are first segmented based on the feature identifiers to obtain a target feature set and at least one boundary feature set; then, connecting the features contained in the target feature set to obtain a target example; and then connecting the features contained in the boundary feature set to obtain at least one boundary instance. Thereby ensuring the accuracy of the example.
Optionally, in the generation process of the target instance, since the target instance relates to the direction of the road, the road may be a curve, so that the target instance needs to be smoothed. Specifically, first, the features included in the target feature set are connected to obtain a route instance, such as a polyline; and then curve fitting is carried out on the route examples to obtain target examples, so that the accuracy of the target examples is improved. For a specific curve fitting process, a cubic B-spline curve fitting algorithm can be adopted, namely, a discrete point group is extracted from a route example; then, pairing discrete points in the discrete point group to determine a corresponding direction vector; further determining control points based on the direction vectors to determine fitting parameters; and performing curve fitting according to the fitting parameters to obtain a target example.
In the process of fitting the cubic B-spline curve, a parameter equation of the curve is defined as:
wherein x, y and z are fitting dimensions; a. b and c are coordinates of the fitting points respectively; and t is more than or equal to 0 and less than or equal to 1.
Then, four adjacent discrete points are arbitrarily extracted from the data to form a discrete point group, pi=(xi,yi,zi) (i-1, …,4) so that a cubic B-spline curve can be determined. The corresponding fitting parameters are:
further, continuously extracting four adjacent discrete points (p) by a window sliding methodi-2,pi-1,pi,pi+1) And calculating two direction vectors after pairing the discrete pointsAndtwo new control points are then calculatedAndand using the newly calculated four points (p)i-1,p′i-1,p′i,pi) And fitting B spline parameters. And finally, repeating the head and tail points of the discrete points to ensure that curve fitting can cover all the discrete points, thereby ensuring the accuracy of curve fitting.
Optionally, it is considered that a situation that the fitted curve does not conform to the real highway pavement due to the accuracy problem of the deep neural network instance segmentation occurs in the process of fitting the curve. In a possible scenario, when a False positive (False positive) condition may occur, adding outliers into curve fitting, which causes outliers in input data of curve fitting to finally cause curve trend errors, for the scenario, fluctuation information of adjacent point locations in a route instance may be determined first; then screening the fluctuation information based on a simplified algorithm to update adjacent point locations; and then adjusting the route instance based on the updated adjacent point location. I.e. a simplified algorithm (e.g. Douglas-Peucker algorithm) is used to simplify the data and reduce the number of outliers caused by false positives.
In another possible scenario, False negative (False negative) conditions may cause data that should be added to the curve fit to be discarded, possibly leading to a broken curve. For the scene, the start and stop point information corresponding to the route instance can be obtained; then determining a target connecting line based on the start point and the stop point information; and updating the route instance according to the intersection condition of the target connecting line and the route tangent, wherein the route tangent is determined based on the route instance. Namely, the relationship between the starting point and the end point connecting line of the similar curves and the tangent line is calculated, and the adjacent curves meeting the geometric constraint condition are connected. For the generation of the target connecting line, simple straight line connection can be used, and a curve extrapolation method can be used for connection. The center line curve calculated by the method can accurately describe the shape and the trend of the main road of the road.
304. A target space is established based on the backbone information and the range information.
In this embodiment, the process of establishing the target space, that is, the process of performing the preliminary reconstruction on the target road, may, for example, first determine the candidate range based on the backbone information, for example, calculate an offset curve of the centerline by the specification of the highway design scheme, and obtain a left curve and a right curve parallel to the centerline curve; the candidate range is then updated according to the range information to establish the target space. The method comprises the steps of obtaining a road edge stone and an isolation zone according to range information, and calculating the parts of the road edge stone, which are close to the left and the right relative to a left curve, of the isolation zone respectively to obtain a target space.
The target road has a branch road and a local widening. It is possible to first determine the local variation region indicated in the range information; then determining the distance information between the local change area and the candidate range based on ray projection; and updating the candidate range according to the distance information to establish a target space. The local change area may be an indication intersection, as shown in fig. 6, and fig. 6 is a scene schematic diagram of another road scene generation method provided in the embodiment of the present application. The figure shows the distance change of the intersection, S1, so that the candidate range can be updated according to the distance change, S1, to establish the target space; for the case that the local variation region indicates local widening, as shown in fig. 7, fig. 7 is a scene diagram of another road scene generation method provided in the embodiment of the present application. The distance change S2 when the road is locally widened is shown in the figure, so that the candidate range is updated according to the distance change S2, thereby establishing the target space.
In particular, the acquisition of the distance variation may be determined by ray casting and intersection in a geometric space. I.e. the area identified as an intersection and a widening of the road, the geometry of this part-area can be reconstructed exactly from the distance values calculated in the identification process. Based on the calculation process, the main road, the fork road and the road widening area can be accurately identified and reconstructed from the collected point cloud.
305. And inserting the detail features into the target space to generate a target road scene.
In this embodiment, after the target space is determined, the target road scene may be generated by further performing detail filling on the generated road basic frame.
Specifically, the insertion of the detail features may be performed based on the target mesh, that is, the width information corresponding to the target road is determined according to the backbone information and the range information; then establishing a target grid in a target space according to the width information and the backbone information; and then calling the detail features to be mapped into the target grid so as to generate the target road scene. The width information is the road width obtained by counting the distance from the boundary line to the center line; and then, carrying out offset calculation on the central line to obtain a curve corresponding to each road part. And then the intersection of the vertical line and each part curve is calculated along the central line, so that the quadrilateral grids of the basic pavement are obtained.
Further, a concrete model of the green belt can be developed through the center line, as shown in fig. 8, fig. 8 is a scene schematic diagram of another road scene generation method provided by the embodiment of the present application. The figure shows a centre line B1 and a green belt B2, i.e. the green belt B2 is filled after determination of the centre line B1 on the basis of the border of the road. It is also possible to obtain a grid or the like of an extended portion such as a side slope, a gutter, or the like by shifting the boundary lane lines. Then, corresponding material balls are respectively manufactured according to the road surface, the lane lines and other parts, calling can be carried out, and texture and material mapping is carried out corresponding to the grids.
Optionally, in order to increase the fidelity of road surface reconstruction, a chapping effect may be added to the material ball of the road surface, and car marks, random dirty spots, and the like may also be drawn along the lane line, as shown in fig. 9, where fig. 9 is a scene diagram of another method for generating a road scene provided in this embodiment of the present application. The figure shows a lane mark C1 added to the road, thereby improving the authenticity of the road.
Further, corresponding car marks can be drawn according to the driving condition of the road, as shown in fig. 10, fig. 10 is a scene schematic diagram of another method for generating a road scene provided in the embodiment of the present application. The drawing shows a vehicle mark D1 in a normal road and a vehicle mark D2 in an emergency lane, and since the vehicle in the emergency lane runs less, the vehicle mark D2 can be lighter than the vehicle mark D1, so that the reality of road simulation is improved. Specifically, for the generation process of the vehicle mark, the driving record information of the target road may be obtained first; then generating driving texture features based on the driving record information; and updating the target road scene according to the driving texture characteristics.
The effects of chapping, car marks, random dirty spots and the like can be added along the lane lines, and the positions and the directions of the effects can be similar to those of the real highway pavement, so that the authenticity of the road is improved.
By combining the embodiment, the target point cloud data corresponding to the target road is collected; then extracting key features in the target point cloud data, wherein the key features correspond to the feature identifiers; clustering the key features according to the feature identification to obtain a target example and at least one boundary example, wherein the target example is used for indicating the backbone information corresponding to the target road, and the boundary example is used for indicating the range information corresponding to the target road; further establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
A specific scenario is described below. Referring to fig. 11, fig. 11 is a flowchart of another road scene generation method according to an embodiment of the present application, where the embodiment of the present application at least includes the following steps:
1101. and controlling the collection vehicle to run on the target expressway and collecting the point cloud data of the laser radar.
In this embodiment, the control of the collection vehicle may be performed by a terminal device, or may be performed by a server for remote control, and the determination of the target highway may be based on a setting of a road, or may be based on a setting of time, for example, a link corresponding to a road measurement performed for 5 hours.
1102. And merging the point cloud data according to the inertial measurement unit, the global navigation satellite system and the calibration information.
In this embodiment, the process of point cloud data merging is similar to step 301 in the embodiment shown in fig. 3, and is not described herein again.
1103. And identifying the road edge stone, the isolation belt and the light barrier based on the point cloud deep neural network.
In this embodiment, the light barrier is a key feature for indicating the road direction; the road edge stones and the isolation zones are key features for indicating the road boundary, and specifically, the key features can also be isolation nets, road rods, guideboards and the like.
1104. And identifying and reconstructing the main lane, the fork and the lane broadside area.
In this embodiment, the main lane is identified based on the target instance and the boundary instance, and the intersection and the lane broadside region are identified by the local features included in the boundary instance, and the distance is adjusted.
1105. And carrying out grid construction, texture coordinate generation and texture mapping to generate the target road scene.
In this embodiment, the process of grid construction and texture coordinate generation is the generation of the target grid and the marking process, and the texture map is the detail feature, so that the target road scene which is accurate and conforms to the actual scene can be obtained.
In particular, the target road scene may be used for simulation of automatic driving, that is, simulating a driving condition of an automatic driving algorithm in the target road scene.
In addition, the target road scene may also act on an automatic driving object, as shown in fig. 12, fig. 12 is a scene schematic diagram of another road scene generation method provided in the embodiment of the present application. The interactive interface on the automatic driving object is shown in the figure, a user can check the surrounding situation of automatic driving through the interactive interface, and compared with real-time camera shooting, a simulated road scene can know a longer distance, namely the road characteristics considered in a panoramic mode, so that the user experience is improved.
The point cloud-based highway reconstruction and scene generation method provided by the invention can greatly reduce the art cost brought by reconstructing and drawing a real scene in automatic driving simulation, and meanwhile, the method is completely based on the acquired original data, so that the problem of data errors caused by artificial modeling is reduced, the reconstructed highway scene is closer to the reality, and the simulation result of the automatic driving vehicle in the environment is more credible.
In order to better implement the above-mentioned aspects of the embodiments of the present application, the following also provides related apparatuses for implementing the above-mentioned aspects. Referring to fig. 13, fig. 13 is a schematic structural diagram of a device for generating a road scene according to an embodiment of the present application, where the generating device 1300 includes:
the acquisition unit 1301 is used for acquiring target point cloud data corresponding to a target road;
an extracting unit 1302, configured to extract a key feature in the target point cloud data, where the key feature corresponds to a feature identifier;
a clustering unit 1303, configured to cluster the key features according to the feature identifiers to obtain a target instance and at least one boundary instance, where the target instance is used to indicate the backbone information corresponding to the target road, and the boundary instance is used to indicate the range information corresponding to the target road;
a generating unit 1304, configured to establish a target space based on the backbone information and the range information;
the generating unit 1304 is further configured to insert a detail feature into the target space to generate a target road scene.
Optionally, in some possible implementation manners of the present application, the collecting unit 1301 is specifically configured to collect segmented point cloud data corresponding to the target road at each time within a preset time period;
the acquisition unit 1301 is specifically configured to convert the segmented point cloud data into a coordinate system of a target time according to a transformation matrix to obtain converted point cloud data;
the acquisition unit 1301 is specifically configured to perform superposition based on the converted point cloud data to obtain the target point cloud data.
Optionally, in some possible implementations of the present application, the extracting unit 1302 is specifically configured to invoke a target neural network based on the target point cloud data, where the target neural network is trained based on training data including training features, the training features include the feature identifiers, and the training data is related to the target point cloud data;
the extracting unit 1302 is specifically configured to input the target point cloud data into a target neural network, so as to extract the key feature.
Optionally, in some possible implementation manners of the present application, the clustering unit 1303 is specifically configured to segment the key feature based on the feature identifier to obtain a target feature set and at least one boundary feature set;
the clustering unit 1303 is specifically configured to connect features included in the target feature set to obtain the target instance;
the clustering unit 1303 is specifically configured to connect features included in the boundary feature set to obtain at least one of the boundary instances.
Optionally, in some possible implementation manners of the present application, the clustering unit 1303 is specifically configured to connect features included in the target feature set to obtain a route instance;
the clustering unit 1303 is specifically configured to perform curve fitting on the route instance to obtain the target instance.
Optionally, in some possible implementations of the present application, the clustering unit 1303 is specifically configured to extract a discrete point group from the route instance;
the clustering unit 1303 is specifically configured to pair the discrete points in the discrete point group to determine corresponding direction vectors;
the clustering unit 1303 is specifically configured to determine a control point based on the direction vector to determine a fitting parameter;
the clustering unit 1303 is specifically configured to perform curve fitting according to the fitting parameters to obtain the target instance.
Optionally, in some possible implementation manners of the present application, the clustering unit 1303 is specifically configured to determine fluctuation information of adjacent point locations in the route instance;
the clustering unit 1303 is specifically configured to filter the fluctuation information based on a simplified algorithm to update the adjacent point locations;
the clustering unit 1303 is specifically configured to adjust the route instance based on the updated adjacent point location.
Optionally, in some possible implementation manners of the present application, the clustering unit 1303 is specifically configured to obtain start and stop point information corresponding to the route instance;
the clustering unit 1303 is specifically configured to determine a target connection line based on the start and stop point information;
the clustering unit 1303 is specifically configured to update the route instance according to the intersection condition of the target connecting line and the route tangent, where the route tangent is determined based on the route instance.
Optionally, in some possible implementations of the present application, the generating unit 1304 is specifically configured to determine a candidate range based on the backbone information;
the generating unit 1304 is specifically configured to update the candidate range according to the range information to establish the target space.
Optionally, in some possible implementations of the present application, the generating unit 1304 is specifically configured to determine a local variation region indicated in the range information;
the generating unit 1304 is specifically configured to determine distance information between the local variation region and the candidate range based on ray casting;
the generating unit 1304 is specifically configured to update the candidate range according to the distance information, so as to establish the target space.
Optionally, in some possible implementation manners of the present application, the generating unit 1304 is specifically configured to determine width information corresponding to a target road according to the backbone information and the range information;
the generating unit 1304 is specifically configured to establish a target grid in the target space according to the width information and the backbone information;
the generating unit 1304 is specifically configured to invoke the detailed feature to map into the target grid, so as to generate the target road scene.
Optionally, in some possible implementation manners of the present application, the generating unit 1304 is specifically configured to obtain driving record information of the target road;
the generating unit 1304 is specifically configured to generate a driving texture feature based on the driving record information;
the generating unit 1304 is specifically configured to update the target road scene according to the driving texture feature.
Acquiring target point cloud data corresponding to a target road; then extracting key features in the target point cloud data, wherein the key features correspond to the feature identifiers; clustering the key features according to the feature identification to obtain a target example and at least one boundary example, wherein the target example is used for indicating the backbone information corresponding to the target road, and the boundary example is used for indicating the range information corresponding to the target road; further establishing a target space based on the backbone information and the range information; and further inserting the detail features into the target space to generate a target road scene. The method and the device realize the automatic generation process of the road scene, and the method and the device can automatically insert details in the road scene due to the fact that the key features corresponding to the target road are adopted to describe the backbone information, the road generation is adjusted by combining the range information, and the accuracy of the road scene generation is improved.
An embodiment of the present application further provides a terminal device, as shown in fig. 14, which is a schematic structural diagram of another terminal device provided in the embodiment of the present application, and for convenience of description, only a portion related to the embodiment of the present application is shown, and details of the specific technology are not disclosed, please refer to a method portion in the embodiment of the present application. The terminal may be any terminal device including a mobile phone, a tablet computer, a Personal Digital Assistant (PDA), a point of sale (POS), a vehicle-mounted computer, and the like, taking the terminal as the mobile phone as an example:
fig. 14 is a block diagram illustrating a partial structure of a mobile phone related to a terminal provided in an embodiment of the present application. Referring to fig. 14, the handset includes: radio Frequency (RF) circuitry 1410, memory 1420, input unit 1430, display unit 1440, sensor 1450, audio circuitry 1460, wireless fidelity (WiFi) module 1470, processor 1480, and power supply 1490. Those skilled in the art will appreciate that the handset configuration shown in fig. 14 is not intended to be limiting and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
The following describes each component of the mobile phone in detail with reference to fig. 14:
The memory 1420 may be used to store software programs and modules, and the processor 1480 executes various functional applications and data processing of the cellular phone by operating the software programs and modules stored in the memory 1420. The memory 1420 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may store data (such as audio data, a phonebook, etc.) created according to the use of the cellular phone, and the like. Further, memory 1420 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.
The input unit 1430 may be used to receive input numeric or character information and generate key signal inputs related to user settings and function control of the cellular phone. In particular, the input unit 1430 may include a touch panel 1431 and other input devices 1432. The touch panel 1431, also referred to as a touch screen, may collect touch operations performed by a user on or near the touch panel 1431 (for example, operations performed by the user on or near the touch panel 1431 using any suitable object or accessory such as a finger or a stylus, and a range of touch operations on the touch panel 1431 with a gap), and drive a corresponding connection device according to a preset program. Alternatively, the touch panel 1431 may include two parts of a touch detection device and a touch controller. The touch detection device detects the touch direction of a user, detects a signal brought by touch operation and transmits the signal to the touch controller; the touch controller receives touch information from the touch sensing device and converts it to touch point coordinates, which are provided to the processor 1480 and can receive and execute commands from the processor 1480. In addition, the touch panel 1431 may be implemented by various types, such as a resistive type, a capacitive type, an infrared ray, and a surface acoustic wave. In addition to the touch panel 1431, the input unit 1430 may also include other input devices 1432. In particular, other input devices 1432 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, and the like.
The display unit 1440 may be used to display information input by or provided to the user and various menus of the mobile phone. The display unit 1440 may include a display panel 1441, and optionally, the display panel 1441 may be configured in the form of a Liquid Crystal Display (LCD), an organic light-emitting diode (OLED), or the like. Further, touch panel 1431 can overlay display panel 1441, and when touch panel 1431 detects a touch operation on or near touch panel 1431, it can transmit to processor 1480 to determine the type of touch event, and then processor 1480 can provide a corresponding visual output on display panel 1441 according to the type of touch event. Although in fig. 14, the touch panel 1431 and the display panel 1441 are two independent components to implement the input and output functions of the mobile phone, in some embodiments, the touch panel 1431 and the display panel 1441 may be integrated to implement the input and output functions of the mobile phone.
The handset may also include at least one sensor 1450, such as light sensors, motion sensors, and other sensors. Specifically, the light sensor may include an ambient light sensor that adjusts the brightness of the display panel 1441 according to the brightness of ambient light, and a proximity sensor that turns off the display panel 1441 and/or the backlight when the mobile phone is moved to the ear. As one of the motion sensors, the accelerometer sensor can detect the magnitude of acceleration in each direction (generally, three axes), can detect the magnitude and direction of gravity when stationary, and can be used for applications of recognizing the posture of a mobile phone (such as horizontal and vertical screen switching, related games, magnetometer posture calibration), vibration recognition related functions (such as pedometer and tapping), and the like; as for other sensors such as a gyroscope, a barometer, a hygrometer, a thermometer, and an infrared sensor, which can be configured on the mobile phone, further description is omitted here.
WiFi belongs to short-distance wireless transmission technology, and the mobile phone can help a user to receive and send e-mails, browse webpages, access streaming media and the like through a WiFi module 1470, and provides wireless broadband internet access for the user. Although fig. 14 shows the WiFi module 1470, it is understood that it does not belong to the essential constitution of the handset and can be omitted entirely as needed within the scope not changing the essence of the invention.
The processor 1480, which is the control center of the mobile phone, connects various parts of the entire mobile phone by using various interfaces and lines, and performs various functions of the mobile phone and processes data by operating or executing software programs and/or modules stored in the memory 1420 and calling data stored in the memory 1420, thereby integrally monitoring the mobile phone. Alternatively, the processor 1480 may include one or more processing units; alternatively, the processor 1480 may integrate an application processor, which handles primarily operating systems, user interfaces, and applications, etc., with a modem processor, which handles primarily wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 1480.
The handset also includes a power supply 1490 (e.g., a battery) that powers the various components, optionally, the power supply may be logically connected to the processor 1480 via a power management system, thereby implementing functions such as managing charging, discharging, and power consumption via the power management system.
Although not shown, the mobile phone may further include a camera, a bluetooth module, etc., which are not described herein.
In the embodiment of the present application, the processor 1480 included in the terminal also has a function of executing the respective steps of the page processing method as described above.
Referring to fig. 15, fig. 15 is a schematic structural diagram of a server provided in this embodiment, where the server 1500 may generate a large difference due to different configurations or performances, and may include one or more Central Processing Units (CPUs) 1522 (e.g., one or more processors) and a memory 1532, and one or more storage media 1530 (e.g., one or more mass storage devices) for storing an application program 1542 or data 1544. Memory 1532 and storage media 1530 may be, among other things, transient or persistent storage. The program stored on the storage medium 1530 may include one or more modules (not shown), each of which may include a series of instruction operations for the server. Still further, a central processor 1522 may be provided in communication with the storage medium 1530, executing a series of instruction operations in the storage medium 1530 on the server 1500.
The server 1500 may also include one or more power supplies 1526, one or more wired or wireless network interfaces 1550, one or more input-output interfaces 1558, and/or one or more operating systems 1541, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.
The steps performed by the management apparatus in the above-described embodiment may be based on the server configuration shown in fig. 15.
An embodiment of the present application further provides a computer-readable storage medium, in which instructions for generating a road scene are stored, and when the instructions are executed on a computer, the computer is enabled to execute the steps performed by the apparatus for generating a road scene in the method described in the foregoing embodiments shown in fig. 2 to 12.
The embodiment of the present application further provides a computer program product including instructions for generating a road scene, which, when run on a computer, causes the computer to perform the steps performed by the road scene generating apparatus in the method described in the foregoing embodiments shown in fig. 2 to 12.
The embodiment of the present application further provides a system for generating a road scene, where the system for generating a road scene may include the generating apparatus in the embodiment described in fig. 13, the terminal device in the embodiment described in fig. 14, or the server described in fig. 15.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a road scene generating apparatus, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.
Claims (15)
1. A method for generating a road scene, comprising:
collecting target point cloud data corresponding to a target road;
extracting key features in the target point cloud data, wherein the key features correspond to feature identifications;
clustering the key features according to the feature identification to obtain a target instance and at least one boundary instance, wherein the target instance is used for indicating the backbone information corresponding to the target road, and the boundary instance is used for indicating the range information corresponding to the target road;
establishing a target space based on the backbone information and the range information;
and inserting detail features into the target space to generate a target road scene.
2. The method of claim 1, wherein the acquiring point cloud data corresponding to the target road comprises:
collecting segmented point cloud data corresponding to each moment of the target road in a preset time period;
converting the segmented point cloud data into a coordinate system of a target moment according to a transformation matrix to obtain converted point cloud data;
and superposing the converted point cloud data to obtain the target point cloud data.
3. The method of claim 1, wherein the extracting key features in the target point cloud data comprises:
calling a target neural network based on the target point cloud data, wherein the target neural network is obtained by training based on training data containing training characteristics, the training characteristics comprise the characteristic identification, and the training data is related to the target point cloud data;
inputting the target point cloud data into a target neural network to extract the key features.
4. The method of claim 1, wherein the clustering the key features according to the feature identifiers to obtain a target instance and at least one boundary instance comprises:
segmenting the key features based on the feature identification to obtain a target feature set and at least one boundary feature set;
connecting the features contained in the target feature set to obtain the target instance;
and connecting the features contained in the boundary feature set to obtain at least one boundary instance.
5. The method according to claim 4, wherein said concatenating the features included in the target feature set to obtain the target instance comprises:
connecting the features contained in the target feature set to obtain a route example;
and performing curve fitting on the route example to obtain the target example.
6. The method of claim 5, wherein said curve fitting said route instance to obtain said target instance comprises:
extracting discrete point groups from the route instance;
pairing discrete points in the discrete point group to determine corresponding direction vectors;
determining control points based on the direction vectors to determine fitting parameters;
and performing curve fitting according to the fitting parameters to obtain the target example.
7. The method of claim 5, further comprising:
determining fluctuation information of adjacent points in the route example;
screening the fluctuation information based on a simplified algorithm to update the adjacent point locations;
adjusting the route instance based on the updated neighboring point locations.
8. The method of claim 5, further comprising:
acquiring start and stop point information corresponding to the route example;
determining a target connecting line based on the start and stop point information;
and updating the route example according to the intersection condition of the target connecting line and the route tangent, wherein the route tangent is determined based on the route example.
9. The method of claim 1, wherein establishing a target space based on the stem information and the range information comprises:
determining a candidate range based on the backbone information;
and updating the candidate range according to the range information to establish the target space.
10. The method of claim 9, wherein the updating the candidate range according to the range information to establish the target space comprises:
determining a local variation region indicated in the range information;
determining distance information of the local variation region and the candidate range based on ray casting;
and updating the candidate range according to the distance information to establish the target space.
11. The method of claim 1, wherein the inserting detail features into the target space to generate a target road scene comprises:
determining width information corresponding to the target road according to the backbone information and the range information;
establishing a target grid in the target space according to the width information and the backbone information;
invoking a detailed feature mapping into the target mesh to generate the target road scene.
12. The method according to any one of claims 1-11, further comprising:
acquiring the running record information of the target road;
generating a driving texture feature based on the driving record information;
and updating the target road scene according to the driving texture features.
13. An apparatus for generating a road scene, comprising:
the acquisition unit is used for acquiring target point cloud data corresponding to a target road;
the extraction unit is used for extracting key features in the target point cloud data, and the key features correspond to feature identifications;
the clustering unit is used for clustering the key features according to the feature identifiers to obtain a target instance and at least one boundary instance, wherein the target instance is used for indicating the backbone information corresponding to the target road, and the boundary instance is used for indicating the range information corresponding to the target road;
a generating unit configured to establish a target space based on the backbone information and the range information;
the generating unit is further configured to insert a detail feature into the target space to generate a target road scene.
14. A computer device, the computer device comprising a processor and a memory:
the memory is used for storing program codes; the processor is configured to execute the method of generating a road scene according to any one of claims 1 to 12 according to instructions in the program code.
15. A computer-readable storage medium having stored therein instructions which, when run on a computer, cause the computer to execute the method of generating a road scene according to any one of the preceding claims 1 to 12.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011095861.4A CN112435333B (en) | 2020-10-14 | 2020-10-14 | Road scene generation method and related device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011095861.4A CN112435333B (en) | 2020-10-14 | 2020-10-14 | Road scene generation method and related device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112435333A true CN112435333A (en) | 2021-03-02 |
CN112435333B CN112435333B (en) | 2022-03-15 |
Family
ID=74690568
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011095861.4A Active CN112435333B (en) | 2020-10-14 | 2020-10-14 | Road scene generation method and related device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112435333B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113362420A (en) * | 2021-06-02 | 2021-09-07 | 北京百度网讯科技有限公司 | Road marking generation method, device, equipment and storage medium |
WO2023050646A1 (en) * | 2021-09-28 | 2023-04-06 | 上海仙途智能科技有限公司 | Device positioning system and method |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108334523A (en) * | 2017-01-20 | 2018-07-27 | 腾讯科技(深圳)有限公司 | The construction method and device of road scene map |
CN108334802A (en) * | 2017-01-20 | 2018-07-27 | 腾讯科技(深圳)有限公司 | The localization method and device of roadway characteristic object |
CN109635641A (en) * | 2018-11-01 | 2019-04-16 | 百度在线网络技术(北京)有限公司 | Determination method, apparatus, equipment and the storage medium of lane boundary line |
CN110598743A (en) * | 2019-08-12 | 2019-12-20 | 北京三快在线科技有限公司 | Target object labeling method and device |
CN110796707A (en) * | 2019-09-10 | 2020-02-14 | 腾讯科技(深圳)有限公司 | Calibration parameter calculation method, calibration parameter calculation device and storage medium |
-
2020
- 2020-10-14 CN CN202011095861.4A patent/CN112435333B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108334523A (en) * | 2017-01-20 | 2018-07-27 | 腾讯科技(深圳)有限公司 | The construction method and device of road scene map |
CN108334802A (en) * | 2017-01-20 | 2018-07-27 | 腾讯科技(深圳)有限公司 | The localization method and device of roadway characteristic object |
CN109635641A (en) * | 2018-11-01 | 2019-04-16 | 百度在线网络技术(北京)有限公司 | Determination method, apparatus, equipment and the storage medium of lane boundary line |
CN110598743A (en) * | 2019-08-12 | 2019-12-20 | 北京三快在线科技有限公司 | Target object labeling method and device |
CN110796707A (en) * | 2019-09-10 | 2020-02-14 | 腾讯科技(深圳)有限公司 | Calibration parameter calculation method, calibration parameter calculation device and storage medium |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113362420A (en) * | 2021-06-02 | 2021-09-07 | 北京百度网讯科技有限公司 | Road marking generation method, device, equipment and storage medium |
CN113362420B (en) * | 2021-06-02 | 2023-08-29 | 北京百度网讯科技有限公司 | Road label generation method, device, equipment and storage medium |
WO2023050646A1 (en) * | 2021-09-28 | 2023-04-06 | 上海仙途智能科技有限公司 | Device positioning system and method |
Also Published As
Publication number | Publication date |
---|---|
CN112435333B (en) | 2022-03-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3944147A1 (en) | Target detection method, model training method, device, apparatus and storage medium | |
EP3951654A1 (en) | Image classification model training method, and image processing method and device | |
CN111598900B (en) | Image region segmentation model training method, segmentation method and device | |
CN112686197B (en) | Data processing method and related device | |
CN110852942B (en) | Model training method, and media information synthesis method and device | |
US20210152751A1 (en) | Model training method, media information synthesis method, and related apparatuses | |
CN112435333B (en) | Road scene generation method and related device | |
CN114092920B (en) | Model training method, image classification method, device and storage medium | |
CN112329725B (en) | Method, device and equipment for identifying elements of road scene and storage medium | |
CN112203115B (en) | Video identification method and related device | |
CN111192341A (en) | Method and device for generating high-precision map, automatic driving equipment and storage medium | |
CN113723378B (en) | Model training method and device, computer equipment and storage medium | |
CN116518960B (en) | Road network updating method, device, electronic equipment and storage medium | |
CN114722937B (en) | Abnormal data detection method and device, electronic equipment and storage medium | |
CN113762044A (en) | Road recognition method, road recognition device, computer equipment and storage medium | |
CN113822263A (en) | Image annotation method and device, computer equipment and storage medium | |
CN113822427A (en) | Model training method, image matching device and storage medium | |
CN114385662A (en) | Road network updating method and device, storage medium and electronic equipment | |
CN115471495B (en) | Model robustness detection method, related device and storage medium | |
CN116665219A (en) | Data processing method and device | |
CN116434173A (en) | Road image detection method, device, electronic equipment and storage medium | |
CN110795994B (en) | Intersection image selection method and device | |
CN113819913A (en) | Path planning method and device, computer equipment and storage medium | |
CN112732446A (en) | Task processing method and device and storage medium | |
CN117911974B (en) | Data processing method, device, equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40040670 Country of ref document: HK |
|
GR01 | Patent grant | ||
GR01 | Patent grant |